Decision-Dominant Logic: Reimagining Value Creation & Capture in the Algorithmic Age (Part 1)
Moving from Service-Dominant Logic to Decision-Dominant Logic to Improve your Competitiveness
Executive Summary
Two decades ago, Vargo and Lusch (2004) challenged the traditional goods-centred paradigm by introducing Service-Dominant (S-D) logic, a relational, co-creative view of value creation. S-D logic reframes exchange as the reciprocal provisioning of service – the application of knowledge and skills for the benefit of others – and depicts value as co-created through interactions among firms, customers, and other actors. This perspective proved influential in marketing and management by shifting focus from tangible product outputs to intangible service processes, collaborative ecosystems, and customer experiences. S-D logic is built on the insight that operant resources (e.g. knowledge and competencies) rather than operand resources (physical goods) are the fundamental source of competitive advantage. By treating customers as co-creators of value rather than passive recipients, S-D logic offered a more dynamic, customer-centric worldview than a Goods-Dominant (G-D) logic that preceded it.
However, as the business landscape undergoes a dramatic transformation driven by big data, machine learning (ML), generative AI, and algorithmic decision-making, it is imperative to ask whether the dominant logic of value creation must evolve once again. We are now in an era where artificial intelligence (AI), data analytics, and algorithms permeate organisational processes. Firms like Amazon, Google, Alibaba, and Uber coordinate value creation and capture not just through service exchanges, but through millions of automated decisions – recommendation algorithms, dynamic pricing engines, real-time logistics optimisations, fraud detection models, and more – executed every minute across their operations. In short, the primary source of competitive advantage appears to be shifting toward the capacity to make superior decisions at scale and speed by leveraging AI and vast data.
I propose Decision-Dominant (D-D) logic as the emerging paradigm to describe this shift. In essence, I argue that decisions – the choices that allocate resources, personalise offerings, set prices, detect risks, and configure operations – are becoming the fundamental unit of value creation and capture in modern firms. In an era of ubiquitous data and powerful algorithms, superior value creation or capture emerges from the quality, speed, and scalability of the decisions made. Firms that can consistently make high-quality decisions quickly and execute them widely (often via automation) hold a substantial advantage over those that cannot.
I label this perspective “decision-dominant” logic to emphasise that intelligent decision processes, powered by analytical and AI capabilities, now dominate value creation and capture much as service interactions did in the S-D logic paradigm.
This article series, written from a strategic management perspective, develops the conceptual foundations of Decision-Dominant logic as a complement – not a replacement – to the existing goods-dominant and service-dominant logics.
In this first article, I review the core principles of Goods-Dominant (G-D) and Service-Dominant (S-D) logic, highlighting their limitations in explaining new mechanisms of value creation and capture in the algorithmic age. Next, drawing on decision science and emerging phenomena like algorithmic management, I articulate why decision-making competence should be elevated to the centre of understanding competitive advantage. I then propose several foundational premises of Decision-Dominant logic, illustrating how it builds upon but also extends beyond G-D and S-D logic. Throughout, I anchor my arguments in established theories – including dominant logic in strategy, dynamic capabilities, organisational learning, organisational information processing, and the resource-based view – to ensure conceptual rigour and connection with extant management theory. I also anticipate potential criticisms and challenges, such as concerns regarding the definitional scope of “decision,” the novelty of this perspective in relation to existing decision theories, and the ethical implications of algorithmic decision systems, and discuss how Decision-Dominant logic addresses these issues. In closing, I argue that Decision-Dominant logic offers a timely and complementary lens for understanding value creation and capture in the algorithmic economy, enriching our theoretical toolkit alongside G-D and S-D logic.
The Algorithmic Age
In barely a decade and a half, “going digital” has become yesterday’s news. The frontier of competition has shifted again—from connecting people and processes to teaching software how to choose. We have entered an algorithmic age, where companies that prosper are those that can let code make millions of tiny, data-driven decisions faster, cheaper, and more accurately than any human team ever could.
Consider three firms that still resemble classic “pipelines,” selling products or services in a linear fashion from supplier to customer. Toyota now streams sensor data from every press and robot on its lines into prediction models that flag a bearing or valve hours before it fails, slashing downtime and repair costs. Walmart feeds two years of sales data, weather maps, and social chatter into forecasting engines that decide, store by store, what to restock each night—keeping shelves full while inventory levels fall. At sea, Maersk lets an AI pilot pick the best speed and course for each vessel, trimming fuel use by double‑digit percentages and avoiding storms long before they form.
On the platform side, the story is the same but louder. Amazon Marketplace rewrites the price tags on millions of items several times a day, nudging demand and margin exactly where it wants them in real time. Uber constantly re‑balances vehicles and riders with dynamic fares that rise or fall every few minutes, matching supply to demand street by street. And every evening Netflix quietly rearranges the rows on your home screen—right down to the artwork—based on what its recommendation models predict will keep you watching.
What unites these examples is not industry, geography or even business model. It is the decision fabric running beneath them: algorithms that sense, decide, act and learn in a perpetual loop. The firm that spins this loop fastest accumulates data sooner, trains better models and widens the gap again—a flywheel of advantage that traditional digital strategies never contemplated. Understanding how that flywheel works, and how to steer it responsibly, is therefore no longer optional; it is the new literacy of competition in the algorithmic age.
In other words, whereas the “digital” era focused on connecting processes, the algorithmic era focuses on deciding—continuously, autonomously and at scale. This section deepens the introductory argument by tracing the economic contours of the algorithmic age and showing why it demands a Decision‑Dominant (D‑D) logic.
The digital push was mainly about linking databases, apps, and people so information could flow end-to-end. It worked: invoices migrated to the cloud, factory machines started pinging maintenance portals, and customers checked parcel locations on their phones. Yet the big calls—how much to produce, when to discount, which shipment to reroute—still waited for a weekly meeting or a manager’s intuition. That gap between always-on data and stop-start human deliberation set the stage for today’s leap from connected processes to autonomous choices.
Three mutually reinforcing shifts are driving the change.
First is sheer data abundance. Analysts now expect the world to create and copy more than 200 zettabytes of data by 2025—roughly enough to stream every movie ever made three billion times. Sensors glued to turbines, clicks poured out by smartphones, and real-time transaction feeds give firms centimetre-level, millisecond-fast visibility they never had in the early cloud era.
Second comes a burst of algorithmic power. Transformer models—the same architecture that powers large language models—and reinforcement-learning pipelines can now spot weak signals, run “what-if” policy experiments in silico, and tune themselves without hand-coded rules. In research labs and on shop floors these systems are stitching together sub-optimal trajectories, learning hierarchical strategies, and gradually taking on decisions once thought too nuanced for code.
The third piece is cheap, elastic compute. Cloud providers rent out GPU clusters by the minute, while new edge-AI kits push the same math onto wind turbines, surveillance cameras, and even smartphones. A predictive-maintenance model that once required a data-centre rack can now sit metres from the asset it protects, analysing signals and acting in real-time.
Together these forces crush the cost of experimentation. A retailer can A/B-test prices every seven minutes, a ship-owner can simulate routes every hour, and a hospital can re-plan staff rosters as admissions ebb and flow—all without summoning a committee. MIT Sloan has dubbed this mindset “algorithmic business thinking” because it treats every activity, from marketing copy to metal cutting, as an optimisation puzzle solvable by software – and even more strategic, assumption-based decisions are increasingly influenced by algorithms. Understanding how these technologies integrate into self-learning decision engines is no longer a side quest; it is the new playbook for gaining an advantage. In the algorithmic age, the firm that learns to choose fastest—not just connect best—writes tomorrow’s competitive rules.
Competing on the Rate of Learning
Digital connectivity once gave companies an edge. Today, that advantage has migrated to the speed and accuracy of the choices those connections allow. Amazon’s repricing engine, for instance, edits roughly 2.5 million product prices every day—fifty times the cadence of its nearest rivals—protecting margin while lifting conversion and consumer surplus in parallel.
Each of those micro‑decisions is a grain of competitive rent that the traditional service lens struggles to explain.
This pattern is now visible across industries. McKinsey’s 2024 Global AI survey finds that 65 per cent of organisations already embed generative AI in at least one core workflow, and the cohort it labels “early adopters” report material EBIT lifts from their faster, data‑rich decision cycles.
In short, firms no longer compete solely on digital presence; they compete on the combined accuracy, adaptability, and speed of their decision-making fabric.
Boston Consulting Group calls this ongoing contest “competing on the rate of learning.” Closed‑loop systems that tie together proprietary data, self‑updating algorithms and automated execution create a flywheel:
More data → sharper predictions → better actions → yet more data.
The winner is not necessarily the firm with the largest balance sheet but the one whose algorithms learn and act the fastest. BCG’s research shows that such “autonomous learning loops” can adjust to market signals in milliseconds, far beyond the bandwidth of hierarchical decision processes.
Equally important, BCG warns that pure technology is not enough. To turn raw learning speed into sustained advantage, firms must re-architect the organisation so that human judgement and machine intelligence specialise on different time horizons—machines optimising the next second to the next quarter, and humans scanning the social, political, and ecological shifts that play out over years. This “human + machine” design transforms the firm into a self‑tuning enterprise whose default state is experimentation and adaptation. In other words, advantage now lies in the compounded accuracy, scalability, adaptability and speed of a firm’s decision fabric - organised by DECISION ARCHITECTS.
Decision-Dominant Logic makes this shift explicit: value is created, captured, and defended through the quality, scalability, and speed of choices, not through the exchange of goods or even services alone. Understanding how to build and govern these algorithmic flywheels, therefore, becomes essential for any executive who wants to write the new rules of competition.
The algorithmic age introduces an outcome–data algorithmic flywheel:
Raw interactions generate behavioural, operational and environmental data.
Algorithms transform data into predictions/recommendations.
Automated execution applies those predictions/recommendations at scale (pricing, routing, content curation).
Outcome feedback—click‑throughs, delivery times, defect rates—feeds the next model training cycle.
Why a Decision‑Dominant Logic Is Inevitable
The foregoing developments suggest three theoretical gaps that a D‑D logic fills:
Unit of analysis: value stems from the pattern and performance of decisions, not from goods or even service interactions alone.
Scalability principle: algorithmic decisions scale nearly cost‑lessly, amplifying marginal returns to data and model improvements.
Learning imperative: continuous model retraining embeds a non‑human learning loop into the firm’s routines, demanding governance mechanisms absent in prior logics.
By foregrounding decision quality, scalability, velocity, and governance as strategic levers, Decision-Dominant logic equips scholars and executives to explain why a retailer’s repricing script, a bank’s credit-risk model, or a hospital’s triage algorithm can be more pivotal to value creation or capture than the frontline service encounter itself. The algorithmic age does not invalidate Service-Dominant insights about co-creation; it magnifies them, because each beneficiary's experience is now co-produced by an invisible ensemble of human and machine decisions. The following section, therefore, elaborates on the development of a Decision‑Dominant logic on the broad and stable shoulders of the good-centred paradigm and the Service-Dominant logic.
From Goods-Dominant to Service-Dominant to Decision-Dominant: Evolving Logics of Value Creation and Capture
Goods-Dominant logic (G-D logic) represents the traditional paradigm that underpinned business thought through the industrial era. In G-D logic, the purpose of economic activity is to make and distribute units of output (goods) that can be sold for profit. Value is seen as embedded in tangible products during manufacturing, and it is “delivered” to customers at the point of sale (value-in-exchange). Customers are primarily viewed as targets or recipients of these goods – often treated as passive operand resources to be segmented and acted upon by the firm. The managerial focus in G-D logic is on efficiently producing standardised goods, controlling production and distribution, and optimising transactions. This goods-centric mental model, while successful in many settings, tends to be static – it assumes the firm creates value, finished at the point of sale, and then consumed by the customer. As a result, G-D logic can miss ongoing value processes and adaptive innovation, often locking firms into product-centric thinking.
Service-Dominant logic (S-D logic), introduced by Vargo and Lusch (2004), was a direct response to the limitations of the goods-centred view. S-D logic posits that service provision, rather than goods, is fundamental to economic exchange. Here, service is defined broadly as the application of specialised competences (knowledge and skills) through deeds, processes, and performances for the benefit of another. All economies are therefore seen as service economies in essence. Several key shifts in perspective distinguish S-D logic from its predecessor:
Intangible resources and operant roles: Instead of focusing on tangible outputs, S-D logic emphasises intangible, operant resources – knowledge, skills, and other dynamic capabilities that act upon other resources – as the primary source of value creation. Goods are still acknowledged, but only as vehicles for service delivery or distribution mechanisms for service provision. For instance, a car (good) embodies applied knowledge and skills (design, engineering, manufacturing services) and only provides value when used (service-in-use) by a driver.
Co-creation of value: Value is not produced unilaterally by the firm and embedded in outputs; instead, value emerges through co-creation in interactions among providers, customers, and other actors. Customers and beneficiaries are viewed as active participants who integrate resources and contribute to value realisation, each from their respective context. In S-D logic, value is always co-created by multiple actors, including the beneficiary (one of its foundational axioms). For example, the value of a healthcare service is co-created by the doctor’s expertise and the patient’s compliance and feedback.
Value-in-use and relational exchange: Because value is co-created in use, value is ultimately determined by the beneficiary’s experience (value-in-use), not just by the producer’s design. This shifts firms towards relational, longer-term engagement with customers. The role of the firm is to offer value propositions and engage in relationships, rather than to deliver final value outcomes unilaterally. S-D logic thereby encourages deep interaction, learning, and adaptation in ecosystems of exchange, in contrast to transactional orientations more embedded in the goods-centred perspective.
By reframing the unit of exchange as service and the locus of value creation as the interaction rather than the transaction, S-D logic provided a more holistic and dynamic worldview of value creation that resonated with the rise of knowledge economies and service industries in the 2000s. It helped explain competitive advantage in businesses where customer experience, relationships, and continuous innovation were paramount. Indeed, S-D logic’s influence has extended beyond marketing into general management, emphasising that all businesses are service businesses at their core (even manufacturing firms depend on design, logistics, and support services). Today, S-D logic – along with related perspectives such as service ecosystems and customer-centricity – remains a crucial lens for understanding value co-creation in networks and platforms.
Why then propose a Decision-Dominant logic?
I argue that the digital revolution and the advent of algorithmically driven processes call for an additional extension of the dominant logic framework. It is not that S-D logic’s principles have become invalid; instead, they are necessary but not sufficient to account for emerging sources of competitive advantage in an AI-powered, data-rich environment. S-D logic focused on interactive processes and operant resources (knowledge), but it did not explicitly theorise the role of automated, optimisation-oriented decision making that today shapes many service interactions. In the current landscape, firms increasingly create value not only by engaging customers in co-creative service relationships, but by orchestrating vast numbers of micro-decisions using algorithms and data. These algorithmic decisions – often invisible to the customer – determine which options are offered, at what price, in what context, and how operations are configured in real-time. While S-D logic acknowledged technology and knowledge as resources, it did not foreground the idea that making better decisions (in design, targeting, pricing, personalisation, etc.) is itself a central strategic objective. I suggest that elevating “decision” to a first-class concept can enhance our understanding of value creation under conditions of massive data availability and AI.
To illustrate the shift: under S-D logic, a retailer like Amazon creates value through service processes (e.g., convenient online shopping, customer service) and co-production with customers (such as reviews, self-service, etc.). Under Decision-Dominant logic, we highlight that Amazon’s competitive edge also critically depends on automated decision engines – for inventory placement, recommendation rankings, and pricing – that operate at lightning speed and huge scale. Amazon’s systems analyse millions of data points and execute algorithmic decisions continuously: its recommendation engine (a decision system) drives an estimated 35% of sales through tailored suggestions, updated in real time; its dynamic pricing algorithms reprice products as often as every 10 minutes based on demand and competition. These capabilities allow Amazon to deliver more relevant choices and better prices to customers (enhancing value-in-use) while also optimising its own outcomes like conversion rates and margins. S-D logic would classify these algorithms as operant resources (applied knowledge), enabling accurate service. Still, D-D logic goes further in treating the decision process itself as the primary locus of value creation or capture. In other words, value is created and captured by making optimal choices, at scale, in addition to (and underlying) the service interactions with customers.
In summary, Decision-Dominant logic builds on the evolution from G-D to S-D logic by proposing that intelligent decision-making is now a critical lens for understanding value creation and capture. Where G-D logic viewed value in goods, and S-D logic viewed value in collaborative service and use, D-D logic views value as emanating from the effectiveness of decisions, often at scale, that integrate data, analytics, and action.
This perspective is timely, given the rise of what some have called the algorithmic economy, in which decision-making power is increasingly being delegated from humans to AI across various industries. In the following sections, I formalise the premises of Decision-Dominant logic and explain its conceptual foundations in more detail.